Large-area soil moisture (SM) data with high resolution and precision are the foundation for the research and application of hydrological and meteorological models, water resource evaluation, agricultural management, and warning of geological disasters. It is still challenging to downscale SM products in complex terrains that require fine spatial details. In this study, SM data from the Soil Moisture Active and Passive (SMAP) satellite were downscaled from 36 to 1 km in the summer and autumn of 2017 in Sichuan Province, China. Genetic-algorithm-optimized backpropagation (GABP) neural network, random forest, and convolutional neural network were applied. A fusion model between SM and longitude, latitude, elevation, slope, aspect, land-cover type, land surface temperature, normalized difference vegetation index, enhanced vegetation index, evapotranspiration, day sequence, and AM/PM was established. After downscaling, the in situ information was fused through a geographical analysis combined with a spatial interpolation to improve the quality of the downscaled SM. The comparative results show that in complex terrains, the GABP neural network better captures the soil moisture variations in both time and space domains. The GDA_Kriging method is able to merge in situ information in the downscaled SM while simultaneously maintaining the dynamic range and spatial details.

The interaction between a squall line and a supercell and its impact on the genesis of a tornado that occurred in Gaoyou, Jiangsu Province, China on 12 June 2020 were analyzed using multi-source observations. The tornado formed as the result of an intensified meso-γ supercell in a favorable large-scale environment. The supercell developed in front of a squall line and slowly intensified after its formation. Due to its small size and weak intensity, the supercell did not produce any severe weather before the approaching of the squall line. As the squall line entered its mature stage with the formation of a well-organized bow echo, the supercell in front of the bow echo began to rapidly intensify and finally led to the tornado touchdown. The analysis of mesoscale and storm-scale wind fields indicated that the bow echo of the approaching squall line modified the kinematic fields near the supercell in such a way that was favorable for the intensification of the supercell. The interaction between the squall line and the supercell may have played a critical role in the occurrence of this tornado.

This study investigates the impact of assimilating conventional weather observations on the wind forecast over the nearshore region of the East China Sea. Multi-level wind measurements in the boundary layer from five masts near the coast were used to verify the numerical model forecasts. Four numerical experiments with a rapid update cycle were performed to forecast the wind field over the masts. The observation shows that the characteristics of the wind field are distinct between the onshore and offshore masts. The numerical forecasts were able to reproduce the main features of the observed wind field both onshore and offshore. However, the wind forecasts of the offshore masts showed larger BIAS and MAE than those onshore. The forecast skill was shown to be sensitive to different weather events and the choice of control variables in the assimilation. The use of new momentum control variables allows a smaller observation-minus-analysis field compared with the traditional control variables, and the resultant wind forecast showed significant improvements. Further tuning of the new control variable scheme showed little improvement of the wind forecast which demonstrates the importance of maintaining the balance between large-scale and small-scale fields in the analysis. The larger forecast error at the offshore masts was likely due to the distribution of conventional observations and the uncertainties in representing the marine boundary layer in numerical models.